階層的可視化手法を用いたアソシエーション分析によるプロファイリング

抄録

In the field of marketing, questionnaires are often carried out and the acquired data are analyzed in order to design a marketing strategy. Association analysis is one of the most powerful techniques for the analysis of questionnaire data. The purpose of this analysis is to extract important rules, e.g.,{Sex=male ⇒ Evaluation of product A=3}, from the data, called association rules. It is an interesting approach to apply the association analysis to profiling in questionnaire data. However, it is difficult for the analyst to grasp the relations between rules and to find interesting rules from a large set of rules. This paper proposes a visualization method of relations between association rules with hierarchical graph structure to support the association analysis. This paper applies the proposed method to an actual questionnaire data on the impression of clothes for men's suits, and shows that the proposed method supports users to grasp relations between rules and to find interesting rules.